About a month ago, a French colleague of mine drew my attention to a french website, the title of which translates to "Private Cossor must be saved". A Cossor DIDS-400 terminal of the mid-1960's was going to end up in the trash unless someone stepped up to rescue it.

About a month ago, a French colleague of mine drew my attention to a french website, the title of which translates to "Private Cossor must be saved". A Cossor DIDS-400 terminal of the mid-1960's was going to end up in the trash unless someone stepped up to rescue it.

As fortune would have it, I had planned a weeks vacation with my family at my parents in France, and it turned out to be only a short detour to pick up the terminal. I asked Jeff to keep the terminal of another month, to which he happily agreed, and last Friday I picked it up.

  • pickup1

A.C. Cossor was a British electronics company, which was acquired by Raytheon in 1961. Around 1965, they produced a terminal called the DIDS-400 which confomed to the - then still under development - ASCII standard.

For its time, the DIDS-400 had a very high refresh rate - 67.5 times per second - and an exceptionally high resolution - over 1,000 display lines. This was achieved by using an analog character generator rather than a digital ROM-based one. Characters are generated through the use of a Brimar XR-1000 monoscope tube (a British equivalent to the Raytheon CK1414 Symbolray tube). A 7373 microsecond pianowire delay line is used as a character buffer.

  • cataract1
  • cataract2
  • cataract3
  • cataract4

The immediately apparent problem with the terminal is the opaque, cracked appearance of the screen. Fortunately, this is just an effect of the degradation of the glue layer between the CRT tube and the front glass (anti glare and implosion protection). Fortunately, the two glass layers separated easily by inserting a flexible plastic spatula, and the glue came right off after that. I'll re-glue the two pieces again later.

  • open1
  • open2
  • open2a
  • open2b
  • open3
  • open3a
  • open3b

Opening the terminal, there are two separated compartments; the top compartment contains the CRT tube and some associated circuitry, as well as the two-piece Best Products (a subsidiary of Cossor) power supply (split into a low voltage and a high voltage part). The bottom compartment contains the monoscope character tube across the front. Behind it, on the left is a card cage with logic boards, and the monoscope control board on the right. The keyboard attaches to this bottom compartment, but can optionally be placed up to 50 feet away from the terminal.

  • cards1
  • cards1a
  • cards1b
  • cards2
  • cards3
  • cards4
  • cards5
  • cards6
  • cards7
  • cards9

For ease of maintenance, the card cage slides out of the chassis completely (with its wiring harness in place). Clever ejector levers (an excentric metal disc with lever) are placed at the back, and I love those wooden tuned pegs used as card guides. The delay line is mounted to the underside of the card cage, and can be accessed by flipping the card cage over.

  • mono1
  • mono2
  • mono3

The magnetic shield around the monoscope tube is the one part of the terminal that is quite rusty and has suffered some damage as a result.

Fortunately, the monoscope tube itself is unscathed. In the closeup, part of the character matrix etched onto the screen is clearly visible.

The monoscope tube is constructed like a small CRT tube with deflector plates, as commonly found in oscilloscopes, but rather than a phosphor coated screen, there is a metal plate (anode) at the end of the tube that the electron beam hits. In front of this plate is the character screen, another plate, which has the character set etched into it (a matrix of 8 by 8 characters). This screen has a voltage on it that's close to the voltage on the anode plate. Where the characters are etched into the screen, the beam is able to penetrate it, and reach the anode. That means that by measuring the current flowing through the anode, one can determine whether the beam hits the etched part of the character or not.

Unlike most other screens or terminals, the screen is not drawn line-by-line, but character by character. Each character is scanned, in 50 lines per character for a 20 character display, on both the display tube (at the coordinates where the selected character needs to appear) and on the monoscreen tube (at the coordinates where the selected character is etched into the character screen); in essence, the character is copied from the monoscope tube to the display tube. This way, the etched screen inside the monoscope tube acts as a very high resolution character ROM.

 

 

Before I can go much further, I am in desperate need of some documentation for this terminal.